e-Learning is used in various learning environments, such as school education and corporate training. In e-Learning, learners set goals and engage with distributed learning courses to achieve them. For example, they might dedicate ample study time, study consistently during spare moments, or tackle difficult areas first, according to their own learning styles.

However, it is difficult for instructors to frequently monitor each learner’s learning style and provide detailed learning support tailored to each style (Problem 1). Additionally, it is difficult for learners to understand what kind of learning style they are using, and it is particularly challenging for them to notice the state change in their learning styles and how they have approached learning in the past (Problem 2).

The purpose of this study is to develop a learning support system for e-Learning that provides learning support tailored to the learner’s learning styles and visualizes the state change in learning styles. Specifically, we focus on clustering and propose a framework for learning support based on learning styles using learning logs. We then develop a learning support system that identifies which learning pattern group each learner belongs to from the learning logs, provides tailored learning support, and visualizes the state change in learning styles.

To solve the above problems, this study establishes three requirements: the ability to identify learners’ learning styles (Requirement 1), the ability to provide learning support tailored to learning styles (Requirement 2), and the ability to help learners become aware of their own growth and provide opportunities to consider how to approach future learning (Requirement 3).

First, we proposed a framework aimed at numerically quantifying data obtained from learning logs and using clustering to simply classify learners' learning styles. This approach allows us to provide learning support and visualize changes in learning styles.

The steps of the framework are as follows.
Step 1) Using learning logs, analyze the characteristics of each cluster through clustering and classify learning styles.
Step 2) Accumulate learning logs from e-Learning.
Step 3) Upon completion of a course, use the learning logs of that course to identify the learner’s learning style using a machine learning model.
Step 4) Use the learning support system to provide tailored learning support based on the identified learning styles and to visualize the state change of learning styles.
Step 5) Repeat steps 2) to 4).

Then, we analyzed learning styles can be classified by clustering learning logs. As a result, we identified three clusters: a group that started just before the deadline, with few study sessions and short study durations, a group that started immediately after the course period began, with few study sessions and short study durations, and a group that started immediately after the course period began and actively engaged in learning.

Finally, we developed a system to provide learning support based on the framework and the clustering results. By using our system, learners are expected to receive learning support that matches their learning styles during their studies. And, by visualizing the changes in learners' learning styles, it is intended that learners will become aware of their own development and approach their learning more proactively.

In the future, we plan to evaluate the effectiveness of our developed system.

Keywords: e-learning, Learning style, Learning log data, Clustering, Visualizing, Learning support system.